Why did Turkish democracy collapse? A political economy account of AKP’s authoritarianism

Why did Turkish democracy collapse?

A political economy account

of AKP’s authoritarianism

Berk Esen

Bilkent University, Turkey

Sebnem Gumuscu

Middlebury College, USA

Abstract

After decades of multiparty politics, Turkey is no longer a democracy. A theory-upending case, the country has descended into a competitive authoritarian regime under the Justice and Development Party (Adalet ve Kalkınma Partisi—AKP), despite rising income and education levels and strong links with the West. What accounts for democratic breakdown in such an unlikely case? Instead of ideological and institutional factors, we offer a political economy account. We contend that the coalitional ties that the AKP forged with businesses and the urban poor through the distribution of public resources has altered the cost of toleration for the party leadership and their dependent clients, while reducing the cost of suppression for incumbents. This new political calculus led to increasing authoritarianism of the AKP government through securitization of dissent, mounting repression, and systematic violation of civil liberties.

Keywords

AKP, clientelism, democratic breakdown, Recep Tayyip Erdog˘an, Turkey

After decades of multiparty politics, Turkey is no longer a democracy. The country has surpassed many well-known cases of democratic breakdown that occurred in the last two decades, including Thailand, Hungary, Ukraine, and Vene-zuela. As in the case of Venezuela, Turkish democracy died slowly through executive aggrandizement led by the Justice and Development Party (Adalet ve Kalkınma Partisi— AKP), which has risen to power in 2002 and established its dominance in Turkish politics through successive elec-toral victories. Starting in 2011, however, the party gradu-ally eroded institutional checks and undermined political rights and civil liberties in the country. After several years of democratic backsliding, Turkey now fails to meet even the basic requirements of an electoral democracy and expe-rienced the sharpest decline over the past decade (Freedom House, 2018).1Elections are now unfair; civil liberties are systematically violated, and the playing field is highly skewed in favor of the incumbent party (Esen and Gumuscu, 2016). We aim to explain Turkey’s democratic breakdown, which defies much of what we know about regime change in comparative politics.

Why did Turkish democracy collapse after six decades of multiparty politics? Why did voters support a political party that undermined accountability and civil liberties? In addressing these questions, we highlight the role of political-economic factors embedded in coalitional politics to explain democratic collapse. Our framework identifies how partisan redistribution serves as a causal mechanism that alters the cost of losing power for both the ruling elite and their supporters. In turn, the cost of suppression is lowered for the ruling elite. This theoretical account hence explains why winners of democratic politics end up under-mining it.

We argue that the interests of the AKP’s ruling coalition is the driving force behind Turkey’s democratic

Paper submitted 29 January 2020; accepted for publication 14 April 2020 Corresponding author:

Sebnem Gumuscu, Department of Political Science, Middlebury College, Middlebury, VT 05753, USA.

Email: sgumuscu@middlebury.edu

Party Politics 1–17

ªThe Author(s) 2020 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/1354068820923722 journals.sagepub.com/home/ppq

breakdown. Specifically, we claim that the party’s cross-class coalition—triangular dependency (TD)—rests on partisan resource allocation toward the economic elites and urban poor. Coalitions on their own do not lead to demo-cratic breakdown though. The causal mechanism behind democratic collapse is the “partisan redistribution” of resources to coalition partners in an increasingly corrupt manner, which entails “reallocation” of favors and cliente-listic goods from former beneficiaries to pro-AKP busi-nessmen and urban poor. Incumbents’ fear of retribution and prosecution increases their cost of losing power through free and fair elections. Their clients also fear change of government because they receive favors via par-tisan redistribution, and they presume that government change could result in loss of such resources to other social groups. Hence, their “fear of future redistribution” of their current benefits under the AKP government drives their cost of toleration up. This collective aversion of the ruling elite and their clients has consolidated the party base around mutual interests, and this popular support substan-tially reduced AKP leaders’ cost of suppression. The party organization played central role in articulating these inter-ests as well as distribution of resources to party clients. As a result, the government could crack down on dissent with tacit support from its constituency. These dynamics— increasing cost of toleration and decreasing cost of suppres-sion triggered by fear of redistribution—in conjunction eroded democratic accountability in the system, ultimately leading to democratic breakdown.

With this framework, we aim to make several contribu-tions to the study of democratic backsliding and collapse. First, our theory unpacks the most common form of dem-ocratic backsliding in the contemporary period, namely executive aggrandizement by a democratically elected gov-ernment (Bermeo, 2016). Democratic backsliding has been widely discussed in the extant literature (Waldner and Lust, 2018) but few studies have indeed accounted for this polit-ical outcome. Although ideational, institutional, and agency-based accounts offer some insights as to why democracies backslide, we assert that these factors work in tandem with economic factors to determine political outcomes.

Second, we adopt Dahl’s (1971) analytical framework of cost of toleration and suppression to explain democratic breakdown. In agreement with Waldner and Lust (2018: 109), we posit that shifts in cost of toleration and suppres-sion provide a causal mechanism that accounts for why freely elected governments start to undermine democratic politics often with tacit or explicit popular support. Yet, this framework remains limited without a broader theoretical commitment, since it does not explain what forces shift political costs of elected officials and why their constitu-encies would agree to a decrease in accountability and competitiveness. We assert that identifying such factors is key in explaining why and when democracies collapse. We

advance this framework by identifying the causal factors that change cost of toleration and suppression and develop a novel theory of democratic collapse that can be applied to a wide range of cases.

Our theory brings back the redistributive model of dem-ocratic collapse, yet in a fundamentally different form. In contrast to the existing model, we explain why it is the winners of democratic politics who undermine it and not the losers. Acemog˘lu and Robinson (2005) and Boix (2003) have in fact highlighted fear of redistribution as a major force behind the changing calculus of political actors that leads to democratic breakdown; yet, their focus was the losers of democratic politics. We posit that it is not the socioeconomic elites that conspire against democracy (i.e. losers) with armed forces as they suggest. Instead, it is the democratically elected government and their partisan clients who benefit from government’s policies that under-mine democracy for they fear the loss of their privileges once the government changes hands. Whereas the afore-mentioned studies explain some cases of democratic break-down through coup d’´etats, our account offers a causal mechanism for democratic backsliding at the hands of popularly elected leaders. As such, our redistributive model better explains the contemporary wave of democratic back-sliding within the contours of an electoral system than the earlier versions with focus on democratic collapse in the 1960s and 1970s in the developing world.

Third, we build a coalitional theory of democratic col-lapse based on a political-economy account. In contrast to scholars who use coalitional theories to explain regime dynamics with social classes at the center of their analysis (Moore, 1966; Rueschemeyer et al., 1992), we find that coalitions that undermine democracy cut across class lines and bring together different socioeconomic groups. The formation of coalitions rests on an array of factors, that is, economic, sociocultural, and ideological, as we discuss below. The triangular coalition noted in this study, for instance, includes both the crony capitalists who benefit from the patronage of the ruling party and the urban poor, often identified in political and sociocultural terms by the government. Moreover, we posit that interests of social classes are not predetermined as many assume but shaped by government policies and changing coalitional dynamics. As such, we agree with Bellin (2000) that social classes are “contingent” democrats, and two factors in particular alter their political calculus: (1) their dependence on the state and (2) fear of losing their privileges (Bellin, 2000; O’Don-nell, 1973). In contrast to O’DonO’Don-nell, who explains the rise of authoritarian regimes through a pro-coup coalition, we show in this article how different social groups with dem-ocratic attitudes (Gumuscu, 2010) end up supporting an authoritarian party once they grow dependent on the gov-ernment. Our model also implies, echoing Bellin (2000), that these groups would rearticulate their interests if or when the costs of this dependence outweigh its benefits.

Finally, our theory bridges an array of social groups and their interests with the individual-level political calcula-tions. As Waldner and Lust (2018) posit, democratic back-sliding (and collapse for that matter) is best explained through such synthetic approaches that bridge “macro-level structures with micro-“macro-level dynamics.” We offer the first synthetic theory of democratic collapse, to our knowl-edge, driven by elected officials and trace the process in a detailed empirical study.

In what follows, we first discuss the nature of regime change in Turkey. Next, we define the puzzle posed by democratic breakdown in Turkey. The third section surveys the extant literature on democratic collapse. We then explain why Turkey has become authoritarian by way of unpacking the cross-class coalition the AKP has built since coming to power. We trace these processes that began in 2011 and provide evidence from three key developments: Gezi Park protests, June and November 2015 elections, and emergency rule established in the aftermath of the failed coup attempt of 2016. We conclude with a discussion of the implications of our argument.

Is Turkey a case of democratic backsliding

or breakdown?

We define democratic breakdown as “reversion to auto-cracy,” which may happen through sudden or slow death of democracy (Waldner and Lust, 2018). Multiparty com-petition in Turkey dates back to 1950, and yet, Turkish democracy was not consolidated before the AKP either. Tutelary powers of the military and the Kemalist bureau-cracy—including the judiciary—in previous decades posed the primary obstacle before democratic consolidation (O¨ zbudun, 2000). The armed forces were not only a source of democratic breakdown as manifested in Turkey’s inter-mittent and promissory coups (Bermeo, 2016), but they also imposed limits on democratic politics via their legal-institutional role, which amounted to virtual veto power over elected officials. And yet, the political playing field was mostly even; elections were free and fair, and there were no major inequalities when it came to parties’ access to resources. The “bounded uncertainty” of democratic rule was well established, even if the bounds had been deli-neated rather narrowly (Schmitter and Karl, 1991).

When the AKP came to power in 2002, Turkey was therefore an electoral democracy. Indeed, Turkey’s scores for political rights and civil liberties improved significantly under the AKP’s first term in power (2002–2007) due to political reforms conducted as part of the European Union (EU) accession process (see Figure 1). The AKP’s viola-tions of civil liberties, however, began as early as its second term in office (2007–2011). During these initial years, although the political system remained open, incumbents targeted specific groups such as Kurdish nationalists, leftist groups, and ultra-secularists while they politicized the

bureaucracy and the judiciary (O¨ zbudun, 2015) and built a pro-government media with the organizational help of the Gu¨len movement.2

After 2010, however, government repression devoured mainstream opposition parties, civil society, and ordinary citizens who mobilized against the AKP rule. The now obvious trend of democratic backsliding gained momen-tum. Indeed, after 2011 Turkey’s civil liberties deterio-rated, while its political rights started to slide back rapidly after 2014 (see Figure 1). Since 2011, the AKP amassed power, eroded checks and balances to capture state institutions, and monopolized power in the hands of its leader, Recep Tayyip Erdog˘an (O¨ nis¸, 2015). As this process unfolded, the electoral playing field tilted in favor of the AKP eroding the competitiveness of elections and rendering each election since 2011 less competitive than the preceding one. This prolonged backsliding eventually culminated in democratic breakdown in a series of devel-opments following the June 2015 parliamentary elections: the AKP government refused to step down despite losing its parliamentary majority in June and called for snap elections in November, established emergency rule after 2016 coup attempt and President Erdog˘an ruled by decree for 2 years, cracked down on dissent and curtailed political rights, and finally, replaced the parliamentary system with executive presidency with weak checks and balances through a refer-endum in 2017 carried out under emergency rule.

Some could argue that the 2019 local election whereby the AKP lost several municipalities marks the vibrancy of Turkish democracy. The reality, however, is quite compli-cated. On the election night, the party refused to concede defeat in Istanbul—the economic bastion of the country and major source of clientelist relations; pressured the elec-toral commission to cancel the elections and secured a rerun on flimsy legal ground—all signs of a competitive authoritarian regime in action (Esen and Gumuscu, 2019). The AKP conceded only after the opposition won the rerun in June 2019 with a 10-point margin. After the election, Erdog˘an government moved to curtail the power of oppo-sition mayors by limiting their financial resources and threatening them with punitive action. More strikingly, the government replaced 24 elected pro-Kurdish mayors with government appointees. These developments once again affirmed the competitive authoritarian nature of the regime, whereby elections are still real, yet the opposition is facing an uneven playing field and has to surpass a much higher electoral bar than the incumbent party to win (Esen and Gumuscu, 2019).

Defining the puzzle: Turkey as a

theory-upending case

Turkey’s recent authoritarian turn under the AKP is highly puzzling: it is a middle-income country with strong ties to the West. The country possesses several structural

conditions—namely, limited natural resources, a weak landowning class, and strong state legacy—that are condu-cive to democratic consolidation (Angrist, 2004; Ross, 2001). Moreover, Turkey’s recent drift into authoritarian-ism directly challenges the primary tenets of the moderni-zation theory, as noted by Sarfati (2017) and Brownlee (2016), warranting an explanation. As Lipset (1959) hypothesized, there is a positive association between eco-nomic development and democracy. Similarly, scholars claim that economic development raises the democratic prospects of a country by creating a strong middle class and a participatory political culture (Boix and Stokes, 2003). Przeworski and Limongi (1997) convincingly demonstrated that democratic regimes are less likely to revert to authoritarian rule in countries with higher levels of economic development.

Even more surprising is the fact that Turkey’s backslid-ing into a competitive authoritarian regime has occurred concurrently with Turkey’s accession talks with the EU, thereby challenging the Europeanization literature. Admit-tedly, the policies of Merkel and Sarkozy governments, not

to mention the EU’s critical stance toward Turkey on the Cyprus issue, have weakened the EU’s leverage over the AKP government as well as the Turkish public support for EU membership (Yılmaz, 2016).3Until the Gezi protests of 2013, however, European governments welcomed the AKP rule as democratic and promoted Turkey as a model case for the rest of the Middle East. Thus, Turkey’s descent into autocracy also opens into question the role of the West as an anchor for democratization (O¨ nis¸ and Bakır, 2007) and challenge Levitsky and Way’s (2010) argument that com-petitive authoritarian regimes with high linkages to the West are expected to democratize in the short run.

Why do democracies break down?

What then explains Turkey’s democratic breakdown? Political institutions, culture, and economic factors are all structural variables that can trigger the collapse of demo-cratic regimes.4Among them, institutional theories focus on constitutional crafting, institutional design, and party systems in their accounts of democratic breakdown. Turkey

24

26

27

28 28

29 29

_{27 27}

28 28 28

26

24

18

16

29

33

38

_{37 37 37}

36 36 36

₃₅

33

₃₂

29 29

20

16

0

5

10

15

20

25

30

35

40

45

50

Polical Rights

Civil Liberes

Figure 1. Political rights and civil liberties in Turkey, 2002–2018. Source: Freedom House Reports 2003–2018. Note: Please refer to the online version of the article to view the figure in colour.

has had relatively strong political institutions, including its century-old parliament and history of holding elections without any significant fraud. Thanks to its parliamentary system, Turkey has also managed to avoid “perils of pre-sidentialism,” which inflates the probability of breakdown (Linz, 1990). Neither the theory of weak party systems (Seawright, 2012) nor the rise of personalistic leaders (Levitsky and Loxton, 2013) can account for democratic collapse in Turkey. Although the AKP’s meteoric rise occurred against the erosion of support for centrist parties amid 2001 economic crisis, the AKP initially adopted mod-erate policies and presented itself as a centrist party (O¨ zbu-dun, 2006). Moreover, as Istanbul’s popular former mayor, Erdog˘an cannot be considered as a populist outsider as claimed by Castaldo (2018) since the AKP cadres were veteran politicians from both the Islamist movement and center-right parties who held political office in the 1980s and 1990s.

Alternatively, cultural factors such as weak civic atti-tudes (Almond and Verba, 1963), self-expressive values (Inglehart and Welzel, 2005), and low levels of social cap-ital (Putnam, 1994) could trigger democratic breakdown. However, the civic culture in Turkey has not regressed over the past decade; in fact, both urbanization and literacy rates have increased in this period. Turkey has a vibrant civil society with close ties to Europe as demonstrated during the Gezi protests in 2013. Support for democracy has been consistently high in the country, and the appeal of radical Islamism has declined over the years, particularly after the AKP came to power (C¸ arkog˘lu and Toprak, 2006), also refuting essentialist accounts of Islam’s incompatibility with democracy (Fish, 2002).

Other scholars have focused on supply-side theories of political culture to emphasize the role of elites in demo-cratic breakdown (Mainwaring and P´erez-Lin˜´an, 2013). Indeed, one could attribute Turkey’s democratic collapse to President Erdog˘an’s authoritarian tendencies (Yilmaz and Bashirov, 2018). And yet, although democracies are often threatened by autocratic leaders, they rarely collapse. Conditions under which such leaders succeed in cases such as Turkey prove to be crucial. Our account explains how Erdog˘an managed to become an undisputed leader within his movement, why he gradually turned to autocratic pol-icies, and most importantly, why voters continued to sup-port the AKP, despite its increasingly authoritarian policies.

Several studies indeed emphasize political-economic factors. For instance, some scholars suggest that in cases with high levels of economic inequality or low capital mobility, elites try to undermine democratic rule due to their fear of redistribution (Acemog˘lu and Robinson, 2005; Boix, 2003). However, democratic politics had already been established for decades at the time the AKP came to power. Moreover, it was not the existing economic elite who conspired against democracy, but it was the

democratically elected government, as we have already explained above. The recent democratic breakdown cannot be attributed to an economic crisis destabilizing the demo-cratic regime either, as suggested by Haggard and Kaufman (1995). For the Turkish economy grew at a rapid pace following the AKP’s rise to power in late 2002 until the end of the decade, resulting in lower unemployment and poverty rates that expanded the party’s electoral base.

Scholarship on Turkish politics has shed some light on how and why the Turkish democracy has broken down. Scholars generally agree that Turkey has descended into a competitive authoritarian regime, without explaining why the AKP specifically has taken an authoritarian turn (Kalaycıog˘lu, 2015; O¨ zbudun, 2015). Similarly, Somer (2016) identifies AKP rule as a new mass-based, personal-ist, and particularistic authoritarianism but does not explain why this authoritarian retreat occurred in the first place.

To explain democratic backsliding, others highlight fac-tors such as the shifting balance of power in Turkish pol-itics in favor of the AKP, the party’s ideology, and sociocultural cleavages (Kubicek, 2016; Ugur, 2017). Alternatively, they discuss the factors that facilitated dem-ocratic backsliding, that is, the ineffectiveness of social opposition, external anchors (in the form of the EU), AKP’s fear and self-confidence evoking a national mandate to rule as they saw fit (O¨ nis¸, 2015). For instance, Akkoyunlu and O¨ ktem (2016) claim that the AKP’s existential insecurity embedded in the Kemalist republic underpins the recent Islamist project of conquering the regime. Accordingly, series of developments in 2007 and 2008—secular popular mobilization against the party during 2007 presidential elections, a memorandum released by the military demand-ing a secular nominee for this position, and the closure case against the party for its anti-secular activities—inflamed AKP’s sense of insecurity and pulled the party in an author-itarian direction.

Although these accounts sound plausible toward explain-ing backslidexplain-ing, it remains unclear, for instance, why the AKP, which survived a strong challenge from the secular Kemalist establishment (i.e. judiciary and the military) through its electoral support, decided to turn authoritarian once such threats were neutralized between 2007 and 2011. More importantly, even if such fears have driven AKP’s authoritarian turn, these scholars do not explain why the AKP constituency supported government’s undemocratic mea-sures after the Kemalist establishment was marginalized.

Indeed, other scholars have explored AKP’s crony and clientelistic ties to businessmen (Esen and Gumuscu, 2018; Gu¨rakar, 2016; Ocaklı, 2018) and the urban poor (Ark-Yıldırım, 2017; (Ark-Yıldırım, 2020; Yo¨ru¨k, 2012), yet they have not specified the impact of such practices on demo-cratic backsliding. Building on these studies, we put forth a causal analysis that connects these processes often treated in isolation in the literature.

TD and democratic breakdown under

the AKP

We assert that political-economic factors embedded in a coalitional theory better explain Turkish democratic col-lapse than any other factor discussed in the aforementioned literature. More specifically, we use a political economy approach that accounts for allocation of resources and changing balances of socioeconomic power within Turkish society under the AKP to explain democratic breakdown.

At the root of the AKP’s democratic backsliding lays extensive capital accumulation and resource allocation for a cross-class coalition composed by the rising economic elite and the urban poor, respectively. Economic liberal-ization of the 1980s empowered a new group of pious Muslim businessmen (Bug˘ra, 1999) while exposing urban poor to increasing precarity (Gu¨lalp, 2001; O¨ nis¸, 1997). The Islamic parties mobilized these social groups against the so-called ruling coalition of secular upper and middle classes around secular-Islamic cleavage (Eligu¨r, 2010). Upon winning local elections in major metropolitan areas in the 1990s, Islamic parties forged what we call a TD among different segments of its constituency at the local level. As the successor to this political tradition, the AKP solidified a cross-class coalition informed by economic interests as well as sociocultural identities, that of Sunni Muslims of Turkish and Kurdish ethnic origins (excluding heterodox Alevi minority, for instance). AKP’s conservative-nationalist worldview that draws on Islamic sentiments and Turkish nationalism (White, 2012) served as the ideational cement of this coalition and reconciled potential intra-class conflicts within (Tug˘al, 2009).

This coalition building, supported by overall macroe-conomic stability the party achieved in power, rested on a partisan allocation of resources to the AKP clients. More

specifically, the relationship between the ruling party and its constituent groups is riddled with redistribution, favor-itism, clientelism, and corruption. In this relationship (shown in Figure 2), pro-AKP business is dependent on the government for capital accumulation through public procurement, construction permits, cheap credit, and tax reliefs. The AKP, in turn, is dependent on business for financial, material, and human resources in the form of campaign contributions, government-friendly media, donations to pro-AKP charities and foundations, and the provision of goods to the urban poor (Esen and Gumuscu, 2018).

The urban poor provides electoral support the govern-ment needs in order to stay in power and injects “democratic legitimacy” to the system despite its increas-ingly undemocratic character. In exchange for their politi-cal support, these poor voters receive selectively distributed social welfare goods, jobs, and charitable goods from the AKP government and pro-AKP foundations.

Yet, this transaction does not occur in a political vacuum. The party organization, under Erdog˘an’s leader-ship, plays a central role in identifying and reaching out to target groups at the very local level. As Baykan states, AKP’s “massive membership organization” is active year-round and “penetrates the remotest corners of the country” (2018: 8). The party’s local branches, hosting more than 1.5 million activists, in its weekly neighborhood meetings coordinate party’s centrally defined strategy, identify those in need, communicate these needs to the higher levels of the party hierarchy, and ensure effective operation of clientelist distribution.5

Equally importantly, the AKP organization evokes fears of redistribution and retribution among government’s cli-ents in case of government turnover. The party elite and ranks spread the message that if the AKP lost power to the

AKP: dependent on businessmen for private resources & media support; on voters for

electoral support

Voters: dependent on the AKP for social services, on businessmen for

charity Business: dependent on the AKP for

contracts/bids/favors/tax relief/access; on urban poor for cheap labor and polical support for

the AKP

Figure 2. Triangular dependency of the AKP regime. AKP: Adalet ve Kalkınma Partisi. Note: Please refer to the online version of the article to view the figure in colour.

secular opposition, party’s supporters will lose their newly gained status. Hence, not only the party helps build a dependent clientele among the business and the urban poor, which we call TD (see Figure 2), but it also alters their political calculus (higher cost of toleration and lower cost of suppression) in tandem with these groups’ changing material interests.

As the AKP’s cost of toleration of opposition increases, the party takes steps to undermine the competitiveness of the regime and decreases the “bounded uncertainty” inher-ent to democracy. As it secures its constituency’s support for its repressive actions, as we will discuss below, the cost of suppression declines along with AKP’s democratic accountability. Consequently, the AKP can further under-mine competitive elections, persecute the opposition, and systematically violate civil liberties without much political cost.

How different is the AKP from previous

governments?

Many parties in power tried to build clientelist relations with their constituencies in Turkey. Yet, the AKP’s clien-telistic network proved to be qualitatively and quantita-tively superior to previous center-right governments’ clientelism (Sayarı, 2014). First, no other political party commanded such electoral dominance that handed the AKP complete control over public resources. Furthermore, these resources have grown rapidly under the AKP rule thanks to

high levels of economic growth and increasing government expenditures as percentage of gross domestic product (GDP) (Figure 3). Both processes have expanded the pool of resources available for transfers from the government to its clients.

This electoral dominance partly stemmed from strong party organization (Baykan, 2018), a historic feature of Isla-mic movements in the country. The AKP is by far Turkey’s strongest party with around 10 million members, many of whom are actively organized at the neighborhood level. No other political party in modern Turkish history matched the level of organizational depth and breadth as the AKP (Hale and O¨ zbudun, 2010). Attesting to this fact is for a decade and a half now the AKP machine has been brilliant at mobilizing support and winning elections (Akdag˘, 2014; Baykan, 2018; Ocaklı, 2017). Meanwhile, in contrast to center-right parties, the AKP evaded centrifugal tendencies through its centra-lized hierarchical party structure (Baykan, 2018) and lack of internal democracy, which for Kumbaracıbas¸ı (2009) was essential in protecting the party from fragmentation. This is not to suggest that the AKP is a monolithic entity devoid of internal splits. Intra-party disagreements when they occurred were successfully suppressed by Erdog˘an as the party turned into what Baykan calls “a personalist mass party.” Similarly, the party rank and file received material and nonmaterial incentives, which have been decisive in the functioning of the party machine.6

Third, liquidation of social welfare state through neolib-eral reforms starting in the 1980s and accelerating under

0 5 10 15 20 25 30 35 40 0 100000 200000 300000 400000 500000 600000 700000 800000 900000 1000000 Millio n s

Turkey GDP (current US$) Turkey Expense (% of GDP)

Figure 3. Turkish economic growth and state expenses, 1960–2018. Source: World Bank Development Indicators. Note: Please refer to the online version of the article to view the figure in colour.

AKP rule after 2002 allowed for much greater space for privatization of public resources and goods, as well as par-ticularistic and politicized social assistance. The AKP gov-ernment carried out unmatched privatization of public assets worth US$62 billion between 2002 and 2017 (com-pared to US$8 billion between 1983 and 2002) in mostly crony fashion, basically redistributing public resources to its supporters (Esen and Gumuscu, 2018). This amounted to 90% of all privatizations since the establishment of the republic in 1923.

Not only the business but also the poor have received their share of AKP’s largesse. Figure 4 shows nominal increase in social benefits distributed to those who are in need yet without social security. Such benefits cover the impoverished in the society including widows and surviving children of the deceased, elderly, disabled, unemployed (without social security), and families with children at school age. The beneficiaries are selected by a local committee composed of centrally appointed gover-nor and other bureaucrats, creating leeway in their deci-sions (Aytac¸, 2014). The available data suggest that previous governments did not establish similar programs or fund them at the level the AKP has done after coming to power in late 2002.

Fourth, the extent of clientelist distribution reached unprecedented levels without disturbing macroeconomic balances, a task in which AKP’s predecessors had repeat-edly failed. One of the reasons why the AKP could sustain such a high level of clientelism with limited fiscal burden is that the party has partly subcontracted social assistance to non-state actors, particularly Islamic charities, philanthro-pic associations, and civil society organizations financed in

no small part by pro-AKP businessmen (Apaydın, 2015; Bozkurt, 2013). It is through this process of delegation and privatization of social spending that the AKP managed to build strong connections among its different clients—the business and the urban poor—through a network of local municipalities and religious charities (Bug˘ra and Candas¸, 2011; Karaman, 2013). These features have set the AKP, and its clients, apart from previous center-right govern-ments and the Kemalist establishment with similar cliente-list tendencies with respect to the extent of redistribution and its sustainability. We now turn to the details of these policies that underlie TD.

Building a dependent business class

The AKP’s organic ties to businessmen constitute one of the primary building blocks of the new regime (Bug˘ra and Savas¸kan, 2014). Following its gradual erosion of rule of law and partisan control of previously autonomous regula-tory organizations (Atiyas, 2012; O¨ zel, 2012), the ruling party has favored a group of crony businesses that support the government and provide the incumbents with private resources and media support. The practice of privatization and public procurement as well as favorable executive dis-cretion in certain sectors such as construction, health, and energy has carved out room for political maneuvering in capital accumulation (Bug˘ra and Savas¸kan, 2014; Gu¨rakar, 2016). Allocation of resources at an unprecedented level, including natural resources, public properties, and state monopolies, in various key sectors has allowed the AKP government to create and nurture a strong business class dependent on the party for lucrative bids, contracts, 5 000 10 000 15 000 20 000 25 000 30 000 35 000 40 000

M

illio

n

s T

L

Figure 4. Social benefits under the AKP government. Source: Turkish Statistical Institute. AKP: Adalet ve Kalkınma Partisi. Note: Please refer to the online version of the article to view the figure in colour.

extraction rights, favorable privatization deals, and other rewards, including tax relief, favorable taxing practices, and immunity from arbitrary tax fines and audits (Esen and Gumuscu, 2018).

Anatolian based, relatively younger enterprises with conservative political preferences primarily benefited from the AKP’s redistributive politics, new procurement prac-tices, privatization schemes, and urban development proj-ects (Hos¸go¨r, 2011; O¨ nis¸, 2011; O¨zcan and Gu¨ndu¨z, 2015). O¨ zcan and Gu¨ndu¨z (2015: 66–67) find that firms with polit-ical connections to the ruling AKP “had abnormal perfor-mance on average over non-connected firms in increase in sales, value added and profit rate under AKP rule.” These favors came at the expense of the established Istanbul-based bourgeoisie with more secular-liberal tendencies and with stronger ties to the Kemalist era.

Public procurement system, privatizations, and public transfer of private capital from dissidents to party support-ers in particular played critical role in the process of build-ing a dependent business class. The Organisation for Economic Co-operation and Development estimates that countries spend around 15% of their GDP for public pro-curement. The relatively large size of the Turkish economy renders public procurement a key method of resource trans-fer (Esen and Gumuscu, 2018). Indeed, Erdog˘an, a month after the election, stated that “he would not leave the con-struction of a 15,000-kilometer long highway to 60 firms” (Sabah, 2002). Thus, Erdog˘an clearly expressed his will to redistribute resources to the rising and loyal businessmen, who formerly had limited access to such resources. To that effect, the AKP government redesigned the procurement system through more than 150 amendments to the public procurement law (Al Jazeera, 2015). As a result, the Public Procurement Agency lost its autonomy (Ercan and Og˘uz, 2006: 652); the state owned enterprises (SOEs) subsidiary companies controlled by municipalities, and public banks on its privatization agenda were taken out of from the pur-view of this law (Bug˘ra and Savas¸kan, 2014: 79); and restricted and negotiated tender methods increasingly replaced open tender methods primarily in the energy, con-struction, and education sectors. Due to these alterations, the government managed to keep US$500 billion (Bug˘ra and Savas¸kan, 2014: 80) worth of public tenders between 2002 and 2014 immune from bureaucratic/judicial review, facilitating resource transfer to its supporters.

The extent of redistribution through these tenders becomes clear when one considers the fact that the share of high-value public procurement increased sixfold between 2004 and 2011 to reach 8.5% of the GDP (Gu¨rakar, 2016: 72–73). In her recent study of 50,000 high-value public pro-curement contracts awarded between 2004 and 2011, Gu¨r-akar found that politically connected firms received 40% of all contracts while “local firms,” mostly with informal con-nections to the AKP (municipalities, local branches, etc.), received another 45% of all awards. The AKP government

also effectively employed privatization as a tool for gaining control of resource allocation to reward politically connected firms in key sectors such as energy and health. As a result, the government has carried out unprecedented scale of pri-vatization—public assets worth US$62 billion—by way of simplifying the bureaucratic procedures and provisions, con-trolling the Privatization Agency and undermining judicial overview over privatization of public assets (Bug˘ra and Savas¸kan, 2014: 82).

The urban poor

In the meantime, the AKP increasingly adopted targeted social assistance programs at both local and national levels. The AKP’s connection with the urban poor rested on social assistance programs and cash transfers for low-income groups primarily through local governments, charities, and civil society organizations. These mechanisms have rein-forced clientelism and cemented ties between pro-AKP businessmen and the mass electorate. As the party redis-tributed resources to the urban poor in the form of condi-tional cash transfers, consumer items, food stamps, and subsidized electricity, urban poor has grown ever more dependent on the AKP government.

As Cammett et al. (2018) show in their study of Turkish central government spending from 2003 to 2014, the AKP strategically distributes economic and social budget expen-ditures across different constituencies depending on their support for the government. In addition, the AKP-controlled metropolitan municipalities have devised several social welfare programs selectively distributed to voters in exchange for their votes (Kesgin, 2012). For instance, the Ankara metropolitan municipality has distrib-uted food, hygiene products, coal, and clothes especially in areas where the party’s vote share is high. The outreach of these programs increased from 37,000 families in 2001 to 400,000 families in 2008. Meanwhile, Istanbul Metropoli-tan Municipality focused on food stamps and cash allow-ances, as shown in Figures 5 and 6. These stamps could be tendered only in select stores, often owned and operated by pro-AKP businessmen and/or companies run by local municipalities. Indeed, a sizable portion of such funds came from entrepreneurs who were asked to make dona-tions to charity funds in exchange for securing a public contract (Bug˘ra and Candas¸, 2011; Eder, 2010).

The AKP government also established a network of patronage at the district level to complement these trans-fers. Social Assistance and Solidarity Foundation (SYDV) is critical in this regard (Eder, 2010) and confirms the political nature of social transfers. For instance, the amount of monthly social assistance provided by the SYDV increased threefold before the 2009 local elections from 124 million Turkish liras (TL) in February 2008 to 317 million TL in February 2009 (Metin, 2011: 197). While in 2013 2,258,734 households received regular and

1,997,306 households received intermittent social assis-tance (Aile ve Sosyal Politikalar Bakanlıg˘ı, 2014), this number increased to nearly 13 million people by 2016 (Gazete Duvar, 2017). For instance, on the eve of the 2015 general elections, the government distributed a one-time allowance amounting to 63 million TL in 44 provinces (Hu¨rriyet, 2015a) and quadrupled its social spending from February 2014 to February 2015 (Hu¨rriyet, 2015b).

Furthermore, distribution of social assistance lacks transparency and is largely discretionary (Bug˘ra and Candas¸, 2011), thus allowing for the ruling party to target electoral districts of critical political importance.7Indeed, the AKP government has effectively used social assistance as an instrument to maximize political support from low-income groups in key constituencies. According to Yo¨ru¨k (2012), for instance, the recipients of government

475362 633089 984989 747031 731070 1180421 1583148 1994515 2006 2007 2008 2009 2010 2011 2012 2013

Figure 5. Number of food stamps distributed by Istanbul Metropolitan Municipality. Note: Please refer to the online version of the article to view the figure in colour.

1379624728 50575 47797 49884 4866055132 56543 75156 99172 133554 147696 169319 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Figure 6. Number of people receiving cash allowances by Istanbul Metropolitan Municipality. Source: Istanbul Metropolitan Municipality Annual Reports.

assistance came disproportionately from the ranks of the Kurdish minority and were concentrated heavily in the urban and metropolitan areas with high levels of migration from the Eastern countryside. Similarly, Aytac¸ (2014) finds that the incumbent (AKP) directs more conditional cash transfers to districts where it is challenged by an ideologi-cally proximate rival.

The AKP also relied on civil society organizations and charities in meeting the needs of low-income groups (Bug˘ra and Candas¸, 2011; Eder, 2010). Islamic values along with material incentives provided by the government played critical role in raising donations for Islamic charities among the ranks of the business. Voluntaristic aspects of this policy framework hallowed out the corporatist welfare system and opened up greater space for political patronage (Eder, 2010; Metin, 2011). Thus, the party turned social welfare into a partisan commodity by creating an electoral constituency dependent on the party for their daily survival. The AKP’s extensive party organization effectively identi-fied those in need whose support could be extracted while excluding opponents from such benefits. The local party branches also coordinated the distribution of benefits in an effective fashion. Hence, the AKP government has entrenched clientelism in its relationship with the electorate and vertically integrated the urban poor to its electoral coalition.

Why and how does the network of dependency fash-ioned by the AKP government pose a threat to democracy? First, it provides the party with unprecedented access to state and private resources, thereby skewing the political playing field in the AKP’s favor. Thanks to TD: (1) the AKP commands an unprecedented degree of private fund-ing and resource accrual from pro-AKP businessmen and (2) as part of their symbiotic relationship with the AKP government, businessmen fund pro-AKP media allowing the party to strongly influence public opinion and restrict freedom of information even in the absence of government censorship, (3) voters give strong support to the ruling AKP in elections and consent for its policies, and (4) due to its electoral dominance, the party acquires control over the state mechanism and its vast resources.

This resource advantage creates an uneven playing field but does not automatically lead to democratic breakdown. Instead, specific mechanisms inherent to TD, which rest on redistribution and reallocation of resources, through cor-ruption, illicit transactions, and favoritism, increase the

cost of toleration for the AKP and its clients. These mechanisms are entrenched in TD and alter the political calculus of the ruling party as well as its supporters to the point where they can no longer afford to lose power since removal from office would result in rent loss, retribution, prosecution, and redistribution of resources to other social groups excluded from the AKP’s TD.

As displayed in Figure 7, the relationship of depen-dency, embedded in clientelism and corruption, has raised the stakes of the democratic game. Thus, the cost of losing power through free and fair elections has substantially increased particularly for the party elites and their depen-dent clients. As the cost of toleration increased for the AKP, the government cracked down on popular protests, delegitimized and criminalized the opposition, securitized dissent, and systematically violated civil liberties. As a result, the competitiveness in Turkish democracy has declined, ultimately triggering democratic breakdown under the AKP rule. At the same time, the collective aver-sion of the pro-AKP cross-class coalition to losing power consolidated the party base and allowed the cost of suppres-sion of dissent for the government to decline. This political calculus manifested itself in unwavering support for anti-democratic measures the AKP took over the years. As such, the AKP supporters readily accepted the government crack-down on popular protests, opposition parties, and dissent within and outside of the party through sustained electoral support and countermobilization at times of existential cri-sis for the ruling party, as we discuss below. This popular support for the government lowered its cost of suppression and led to erosion of vertical accountability in the political system. Eventually, Turkish democracy has collapsed.

Tracing democratic breakdown in Turkey

Measuring cost of toleration and suppression in a precise manner is a difficult task, yet we can use proxies to estimate these costs for the AKP elite and their supporters. To that effect, we focus on AKP elite’s statements and public opin-ion surveys to identify their political calculus. The AKP supporters’ view of the opposition reflects their cost of toleration, while their perception of the government’s repressive measures against dissidents shows the cost of suppression for the government.

To demonstrate how these costs translate into demo-cratic collapse, we analyze critical developments such as Cross-class coalions (Triangular Dependency) redistribuon from former beneficiaries corrupon/clien telism/unequal access to public and private resources stakes of the democrac game change cost of toleraon increases cost of suppression decreases the AKP systemacally violates civil liberes, harasses the opposion accountability and compeon erodes democrac breakdown

Figure 7. Causal mechanism of democratic breakdown in Turkey. Note: Please refer to the online version of the article to view the figure in colour.

the Gezi protests and emergency rule that deepened the AKP’s sense of threat by way of popular mobilization and coup attempt, respectively. These events indicate the rising cost of toleration and decreasing cost of suppression for the government: Gezi protests marked the most extensive and systematic violation of civil liberties under the AKP rule since 2002, whereas the emergency rule after the coup attempt in July 2016 institutionalized democratic break-down and led to the transition to executive presidency in 2017.

In May 2013, the AKP government moved to raze the Gezi Park in downtown Istanbul to build a new shopping mall. Activists started a sit-in in the park to stop the con-struction. When the police brutally removed them from the park, nationwide demonstrations erupted challenging the AKP government and its urban development projects cater-ing to government cronies. In response, Prime Minister Erdog˘an called protesters as coup-supporters, foreign agents, and looters. The police cracked down on protests killing 11 protesters and injuring thousands with impunity. Despite this heavy use of force by the police, elite and mass defection in the AKP remained quite limited. Reflecting the low cost of suppression, Erdog˘an openly expressed his sup-port for police brutality and commended their action as legendary and heroic.

The AKP constituency shared Erdog˘an’s view to a large extent. In fact, as Aytac¸ et al. find, “ . . . the public’s reac-tions to the Gezi Park protests and to the government’s handling of them were sharply structured by polarized socio-religious and party affinities; virtually no govern-ment supporters joined the Gezi park protests in Istanbul” (Aytac¸ et al., 2017). The authors conclude that “ . . . what gave Erdog˘an the freedom to respond to the Gezi Park protesters so harshly was the near-certainty that repression would have no serious electoral repercussions . . . .” (Aytac¸ et al., 2017: 66). Ninety-two percent of AKP supporters perceived Gezi protests as a conspiracy against the govern-ment, while only 8% claimed that they were an expression of democratic rights and freedoms (KONDA, 2018). Eighty-one percent of AKP supporters thought that Gezi protests were initiated by foreign powers to destroy the AKP government (Erdog˘an, 2018).

Thanks to popular support for its repressive measures, the AKP government securitized dissent by systematically harassing government critics in civil society and crimina-lizing protest behavior. The government also adopted a new National Security Council (MGK) document and passed “the Internal Security Package” in March 2015. While the MGK document identified popular protests as the main threat to national security, the new bill “granted broad dis-cretionary powers to the law enforcement agencies without adequate judicial or independent parliamentary oversight” (European Commission, 2015: 62). Such measures drove AKP’s cost of suppression further down.

The party has undermined the legitimacy of opposition parties in this period. Its leaders consistently vilified the opposition as disrespectful and hostile to the interests of the “people” and equated dissent with terrorism. According to Erdog˘an, the AKP (and its current ally, the nationalist Milliyetc¸i Hareket Partisi–MHP) is the only legitimate political actor that deserves political support. All other political actors are enemies of the people. The AKP leaders increasingly treat oppositional activity as illegitimate actions of “evil” conspirators and tantamount to coups. Recently, Numan Kurtulmus¸, AKP’s deputy chairman, for instance, claimed that an “evil” oppositional bloc tried to oust Erdog˘an in the Gezi Park protests, December 2013 graft probe, the coup attempt in July 2016, and again in the March 2019 local elections (Yeni Akit, 2019).

Interestingly, on numerous occasions, the AKP elite expressed fears of prosecution and retribution if they lost power. An AKP MP cautioned party supporters that if the AKP lost the elections, its constituency, including mer-chants, covered women, small shopkeepers, would suffer (Yeni Asya, 2019). Again, on the campaign trail, an AKP mayor claimed that if the ruling party lost in the local elections, they would be executed or killed in a brutal man-ner (Yenic¸ag˘, 2019). These statements clearly relay the fear that the government actors associate with the loss of power. Such fear is certainly incompatible with democracy, a sys-tem based on peaceful alternation of power, yet also indi-cative of high stakes for the AKP elite. Fueling such fears, we posit, is the unprecedented level of government corrup-tion. A graft probe in late 2013 revealed the illegal exchanges between several AKP ministers and crony busi-nessmen, while Court of Accounts regularly noted such irregular transactions in its annual reports, to no avail, since the government increasingly curtailed judicial oversight since 2010.

The AKP voters seem to agree for the most part with the party elite that the opposition is not an essential part of democratic politics but rather a hindrance. Nearly 75% of AKP supporters think that the opposition slows down the government’s achievements by intervening in its policies (Erdog˘an, 2018). The AKP voters also lent support for social media screenings and systematic harassment of crit-ics: 79% of respondents among AKP supporters in 2017 claimed that they were fine with state-imposed restrictions over the Internet and social media such as Twitter and Facebook, while 58% believed that social media is the biggest menace for societies. Similarly, 76% stated that the government could impose media blackout after certain incidents to maintain social order (KONDA, 2018). These attitudes do not reveal only the high cost of toleration for oppositional activities and freedom of expression among AKP supporters, but also their high tolerance for suppression.

After successfully eliminating challenges posed by pop-ular mobilization and elections, the AKP faced a

nondemocratic threat to its dominance in July 2016 in the form of a putsch. The government with support from the opposition parties managed to thwart the coup. Yet, Erdo-g˘an used the putsch as a pretext to extend his presidential powers and curtail civil liberties and political rights. Nota-bly, in this period the parliament lifted the legal immunity of MPs that culminated in the imprisonment of several pro-Kurdish deputies. The state of emergency also resulted in the purge of 150,000 public sector employees, confiscation of countless firms, the arrest of 60,000 people, including scores of journalists and human rights advocates, and the closure of 1500 civil society organizations along with media organs and educational institutions by the end of 2017 (Freedom House, 2018). Revealing the government’s low cost of suppression, the AKP constituency supported these repressive measures to a great extent. In 2017, for instance, 66% of AKP voters deemed the purge of civil servants as appropriate; 89% supported the confiscation of certain firms, while 91% thought it was rightful to close down some educational institutions. Sixty-two percent also claimed that the government could restrict political rights and civil liberties as part of its counterterrorism and anti-criminal measures (KONDA, 2018).

We argue that the political cost function of the AKP supporters changed due to their increasing dependence on the ruling party. The fear of losing the material gains they acquired under the AKP rule has increased their intolerance for the opposition. The AKP voters’ assessment of their relative status in society under the AKP rule supports our argument. In a survey conducted in 2015, 64% of AKP supporters stated that people in their “group” has gained greater say in the country in the past 5 years; 58% believed that other groups envied them; and 47% thought that they had benefited more compared to other groups in material terms in the past 5 years (Erdog˘an, 2016). Interestingly, in a similar survey conducted last year, MHP voters, whose party recently allied with the AKP, registered a dramatic increase in their favorable responses (Erdog˘an, 2018). Sev-eral interviews with lower income families in Istanbul’s districts conducted by one of the authors confirmed this finding.8When respondents were asked why they supported the AKP, many claimed that they enjoyed substantial mate-rial benefits thanks to the AKP policies and programs.

Consequently, AKP’s electoral support did not waver despite its increasingly authoritarian practices, the party won the 2014 local, 2014 and 2018 presidential elections, November 2015 and June 2018 national elections, and the 2017 constitutional referendum. These electoral victories allowed the AKP to continue violation of civil liberties, crackdown on dissent, and securitize oppositional politics.

Conclusion

While there is a rich debate on democratic collapse, few scholars explain why democracies regress. To address this

gap in the literature, we offer a synthetic approach centered on coalitions. Our analysis highlights political and eco-nomic factors that incentivize the ruling elites’ avoidance to loss of power, while focusing on social forces to explain why this tendency eventually leads to regime breakdown. This interactive account allows us to explain the behavior of both ruling parties and their supporters.

This article hence explains why Turkish democracy col-lapsed after six decades of multiparty politics. We argue that the ruling party created a highly partisan base among economic elites and urban poor through targeted resource distribution. Accordingly, this TD built by the AKP turned into a virtuous cycle whereby the party, its affiliated busi-nesses, and its voters have grown into mutual interdepen-dence. Given that the flow of resources and privileges depended on the AKP’s continued control of the state appa-ratus, both groups became increasingly partisan and fearful of redistribution and retribution if the AKP were to lose power. This mechanism altered “the cost of tolerance” and “the cost of suppression” for party leadership as well as the AKP’s cross-class coalition by raising the stakes of democ-racy. As such, our analysis explains not only the AKP’s reluctance to step down through democratic means but also its constituents’ willingness to tolerate democratic breakdown.

Although the ruling party retained its policy of reward-ing its elite and popular supporters, the makeup of the AKP’s coalition varied over time. With the personalization of power and rising authoritarianism under Erdog˘an’s rule, the need for appealing to voters and businessmen decreased. After more than 15 years in power, the party’s dependency on the business for private resources has decreased in comparison to the dependence of economic elites on the AKP. And yet, the AKP leadership still turns toward pro-government businesses for material support for its projects. Moreover, the ruling party is still heavily dependent on its voters for electoral support to remain in power. Regardless, the AKP’s cross-class coalition nar-rowed with many former government supporters excluded, either due to concentration of corruption at the top or polit-ical rifts. Under certain circumstances, such disintegration may even trigger re-democratization.

For instance, an economic shock that severely limits incumbent’s ability to distribute resources could alter polit-ical calculus of regime’s clients and lead to disintegration in the ruling coalition. After the recent economic downturn, the Erdog˘an administration have had trouble in addressing the needs of his popular base and even faced criticism from the business. When met with effective coordination among the opposition parties, the economic crisis led the ruling party to suffer an electoral setback in the 2019 local elections (Esen and Gumuscu, 2019). Indeed, in accordance with our account, the party refused to concede in Istanbul since the municipal government generated rents through

construction permits, land allocation, municipal-run compa-nies, and local taxes that the AKP elites did not want to lose. Our study thus reaffirms the centrality of resource distribution for authoritarian regimes. In contrast to Greene’s (2010) expectations, when elites turn neoliberal-ism into crony capitalneoliberal-ism instead of well-functioning free markets, they may doom democracies and stabilize author-itarian politics instead of undermining hegemonic parties. The article also complements earlier studies that attrib-uted the moderation of political Islam to the rise of “devout bourgeoisie” who had vested interests in economic liberal-ism and democracy (Gumuscu, 2010). While the Anatolian-based businessmen had a democratizing impact on the Islamist movement in the 1990s, the AKP managed to turn this class actor into dependent clients after assuming power through state patronage. Reminiscent of Bellin’s (2000) contingent democrats, these pious businessmen after integrating into the AKP’s ruling coalition prioritized their interests over democratic commitments. Despite growing political instability and societal opposition, large segments of these businessmen and voters supported the AKP government and its increasingly authoritarian practices.

The political economy model we offer in this article may find relevance in other cases such as Hungary, Brazil, Venezuela, Poland, and South Africa, where ruling parties have cultivated organic ties with a pro-government busi-ness class and voter base. In cases where such ties alter the cost of toleration and suppression for the ruling parties, we would predict a trend of democratic regression similar to what has occurred in Turkey. We claim that our argument has higher explanatory value in other regions that experi-enced the emergence of strong leaders who captured the state apparatus for partisan distribution with great effect. A better understanding of the causes of democratic break-down can shed greater light on what is clearly a wave of democratic reversal at a global level (Diamond, 2015). Acknowledgments

The authors would like to thank Razan Jabari, Alice Lo˜hmus, and Hazan Sucu for their valuable research or technical assistance and Samuel Greene, Sahar Aziz, Ziya O¨ nis¸, Andreas Schedler, Zeki Sarıgil, Nora Fisher Onar, David Waldner, and participants at the interdisciplinary workshop on the “Political Economy of Contem-porary Turkey” (2015), the inaugural symposium “Consortium of European Symposia on Turkey” (2015), the “Rise of Competitive Electoral Authoritarian Regimes” panel at APSA 2015, “Contesting Autocracy in the Middle East” Panel at MPSA (2017), and “Empirical Studies in Political Analysis” workshop in Turkey (2018) for their comments on earlier drafts of this article.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sebnem Gumuscu https://orcid.org/0000-0002-9180-421X

Notes

1. Freedom House defines electoral democracies as those coun-tries that receive at least 7 in the electoral process subcategory and at least 20 in political rights. In 2016, Turkey received 8 for its electoral process and 18 ( 6 points change) in its overall political rights scores (Freedom House, 2018).

2. The Gu¨len movement was established in 1966 to raise a “golden generation” of pious Muslims with a strong sense of solidarity and discipline. Its leader, Fethullah Gu¨len, aimed to colonize the state bureaucracy with his followers, primarily through manipulation of bureaucratic recruitment processes. 3. We thank the referees for raising this point.

4. For a comprehensive review of the literature, see Waldner and Lust (2018).

5. Interviews conducted with six local party officials of the AKP in Umraniye and Kartal in 2011. Interviewees include heads of the district branch, officials in charge of social programs in the municipality, and members of neighborhood branches. Also see Erdog˘an (2016), Eligu¨r (2010), and Baykan (2018). 6. For more on ideational and collective incentives the party

leadership had to offer to its ranks, see Baykan (2018: Ch. 5). 7. Overall, 32 million Turkish liras (TL) was spent in Istanbul, 17 million TL in Diyarbakır, and 11 million TL in S¸anlıurfa (Metin, 2011: 197).

8. Fieldwork conducted by one of the authors in two districts of Istanbul in 2011 and 2012 and funded by the Istanbul Policy Center. In total, 119 respondents were interviewed by the author and Erdem Aytac¸ in six neighborhoods in Umraniye and Kartal.

References

Acemog˘lu D and Robinson JA (2005) Economic Origins of Dic-tatorship and Democracy. Cambridge: Cambridge University Press.

Aile ve Sosyal Politikalar Bakanlıg˘ı (2014) 2013 Yılı Faaliyet Raporu. Available at: https://ailevecalisma.gov.tr/media/3 6155/2013-yili-faaliyet-raporu.pdf (accessed 9 February 2017). Akdag˘ G (2014) Ethnicity and Elections in Turkey: Party Politics and the Mobilization of Swing Voters. Abingdon: Routledge. Akkoyunlu K and O¨ ktem K (2016) Existential insecurity and the

making of a weak authoritarian regime in Turkey. Southeast European and Black Sea Studies 16(4): 505–527.

Al Jazeera (2015) Kamu Ihale Kanunu yine degisiyor. 15 January. Available at: http://www.aljazeera.com.tr/haber/kamu-ihale-kanunu-yine-degisiyor (accessed 10 June 2015).

Almond GA and Verba S (1963) The Civic Culture: Political Attitudes and Democracy in Five Nations. Princeton: Prince-ton University Press.

Angrist MP (2004) Party systems and regime formation in the modern Middle East: explaining Turkish exceptionalism. Comparative Politics 36(2): 229–249.

Apaydın F (2015) Financialization and the push for non-state social service provision: philanthropic activities of Islamic and conventional banks in Turkey. Forum for Development Studies 42(3): 441–465.

Ark-Yıldırım C (2017) Political parties and grassroots clientelist strategies in urban Turkey: one neighbourhood at a time. South European Society and Politics 22(4): 473–490.

Atiyas I (2012) Economic institutions and institutional change in Turkey during the neoliberal era. New Perspectives on Turkey 47(1): 57–81.

Aytac¸ SE (2014) Distributive politics in a multiparty system: the conditional cash transfer program in Turkey. Comparative Political Studies 47(9): 1211–1237.

Aytac¸ SE, Schiumerini L and Stokes S (2017) Protests and repression in new democracies. Perspectives on Politics 15(1): 62–82.

Baykan TS (2018) The Justice and Development Party in Turkey: Populism, Personalism, Organization. Cambridge: Cambridge University Press.

Bellin E (2000) Contingent democrats: industrialists, labor, and democratization in late-developing countries. World Politics 52(2): 175–205.

Bermeo N (2016) On democratic backsliding. Journal of Democ-racy 27(1): 5–19.

Boix C (2003) Democracy and Redistribution. Cambridge: Cam-bridge University Press.

Boix C and Stokes SC (2003) Endogenous democratization. World Politics 55(4): 517–549.

Bozkurt U (2013) Neoliberalism with a human face: making sense of the Justice and Development Party’s neoliberal populism in Turkey. Science & Society 77(3): 372–396.

Brownlee J (2016) Why Turkey’s authoritarian descent shakes up democratic theory. Washington Post. Available at: https:// www.washingtonpost.com/news/monkey-cage/wp/2016/03/ 23/why-turkeys-authoritarian-descent-shakes-up-democratic-theory/ (accessed 10 April 2016).

Bug˘ra A (1999) Islam in Economic Organizations. TESEV Report.

Bug˘ra A and Candas¸ A (2011) Change and continuity under an eclectic social security regime: the case of Turkey. Middle Eastern Studies 47(3): 515–528.

Bug˘ra A and Savas¸kan O (2014) New Capitalism in Turkey: The Relationship Between Politics, Religion and Business. Chel-tenham: Edward Elgar Publishing.

Cammett M, Luca D and Sergenti E (2018) Unfair play: the pol-itics of Turkey’s central government spending under AKP rule. Available at: https://www.researchgate.net/publication/3 23800418_Unfair_play_The_politics_of_Turkey%27s_cen tral_government_spending_under_AKP_rule (accessed 24 September 2018).

C¸ arkog˘lu A and Toprak B (2006) Deg˘is¸en Tu¨rkiye’de Din Toplum ve Siyaset. _Istanbul: TESEV Yayını.

Castaldo A (2018) Populism and competitive authoritarianism in Turkey. Southeast European and Black Sea Studies 18 (4): 467–487.

Dahl R (1971) Polyarchy: Participation and Opposition. New Haven: Yale University Press.

Diamond L (2015) Facing up to the democratic recession. Journal of Democracy 26(1): 141–155.

Eder M (2010) Retreating state? Political economy of welfare regime change in Turkey. Middle East Law and Governance 2(2): 152–184.

Eligu¨r B (2010) The Mobilization of Political Islam in Turkey. Cambridge: Cambridge University Press.

Ercan F and Og˘uz S¸ (2006) Rescaling as a class relationship and process: the case of public procurement law in Turkey. Polit-ical Geography 25(6): 641–656.

Erdog˘an E (2016) Tu¨rkiye’de Kutuplas¸manın Boyutları Ara-s¸tırması. Available at: http://kssd.org/site/dl/uploads/Kutu plas¸ma-Aras¸tırması-Sonuc¸ları.pdf (accessed 8 January 2019).

Erdog˘an E (2018) Tu¨rkiye’de Kutuplas¸manın Boyutları Ara-s¸tırması. Available at: https://goc.bilgi.edu.tr/media/uploads/ 2018/02/05/bilgi-goc-merkezi-kutuplasmanin-boyutlari-2017-sunum.pdf (accessed 8 January 2019).

Esen B and Gumuscu S (2016) Rising competitive authoritarian-ism in Turkey. Third World Quarterly 37(9): 1581–1606. Esen B and Gumuscu S (2018) Building a competitive

author-itarian regime: state–business relations in the AKP’s Tur-key. Journal of Balkan and Near Eastern Studies 20(4): 349–372.

Esen B and Gumuscu S (2019) Killing competitive authoritarian-ism softly: the 2019 local elections in Turkey. South European Society and Politics 24(3): 317–342.

European Commission (2015) Turkey 2015 Progress Report. Available at: https://ec.europa.eu/neighbourhoodenlarge ment/sites/near/files/pdf/key_documents/2015/20151110_ report_turkey.pdf (accessed 24 January 2017).

Fish MS (2002) Islam and authoritarianism. World Politics 55(1): 4–37.

Freedom House (2018) Freedom in the World 2018. Available at: https://freedomhouse.org/report/freedom-world/freedom-world-2018 (accessed 25 August 2019).

Gazete Duvar (2017) 13 milyon kis¸i sosyal yardımlarla yas¸ıyor. 22 December. Available at: https://www.gazeteduvar.com.tr/ gundem/2017/12/22/13-milyon-kisi-sosyal-yardimlarla-yasiyor (accessed 12 January 2018).

Greene KF (2010) The political economy of authoritarian single-party dominance. Comparative Political Studies 43(7): 807– 834.

Gu¨lalp H (2001) Globalization and political Islam: the social bases of Turkey’s Welfare Party. International Journal of Middle East Studies 33(3): 433–448.

Gumuscu S (2010) Class, status, and party: the changing face of political Islam in Turkey and Egypt. Comparative Political Studies 43(7): 835–861.